American police departments ... are paying hundreds of thousands of dollars for an unproven and secretive technology that uses AI-generated online personas designed to interact with and collect intelligence on "college protesters," "radicalized" political activists, suspected drug and human traffickers ... with the hopes of generating evidence that can be used against them. Massive Blue, the New Yorkbased company that is selling police departments this technology, calls its product Overwatch, which it markets as an "AI-powered force multiplier for public safety" that "deploys lifelike virtual agents, which infiltrate and engage criminal networks across various channels." 404 Media obtained a presentation showing some of these AI characters. These include a "radicalized AI" "protest persona," which poses as a 36-year-old divorced woman who is lonely, has no children, is interested in baking, activism, and "body positivity." Other personas are a 14-year-old boy "child trafficking AI persona," an "AI pimp persona," "college protestor," "external recruiter for protests," "escorts," and "juveniles." After Overwatch scans open social media channels for potential suspects, these AI personas can also communicate with suspects over text, Discord, and other messaging services. The documents we obtained don't explain how Massive Blue determines who is a potential suspect based on their social media activity. "This idea of having an AI pretending to be somebody, a youth looking for pedophiles to talk online, or somebody who is a fake terrorist, is an idea that goes back a long time," Dave Maass, who studies border surveillance technologies for the Electronic Frontier Foundation. "The problem with all these things is that these are ill-defined problems. What problem are they actually trying to solve? One version of the AI persona is an escort. I'm not concerned about escorts. I'm not concerned about college protesters. What is it effective at, violating protesters' First Amendment rights?"